Real-time estimation and prediction for pandemic A/H1N1(2009) in Japan.

The Infectious Disease Surveillance Center, National Institute of Infectious Diseases, Japan, has developed a fully automatic system for daily reporting of ILI (influenza-like illness) patients prescribed Oseltamivir and Zanamivir. The system collected data on the number of prescriptions from approximately 3,350 pharmacies, which account for almost 7% of all pharmacies in Japan, and published the number online on a daily basis, except for Sundays. On the basis of these data, we further estimated R (v) and predicted its course every Monday using a very simple SIR model with data from previous studies. This paper summarizes our real-time estimation and prediction of ILI patients in the 2009 influenza A/H1N1 pandemic. The estimate on November 29 resulted in an R (v) of 1.72 (95%CI [1.69, 1.75]). The model predicted that the peak of the epidemic would be reached on December 23 2009 [December 14, 2009, January 2, 2010] with an estimated number of ILI patients of 227,000 [193,000, 262,000]. The cumulative number of ILI patients over the period would be as high as 17.8% [16.6, 19.0%] of the total population of Japan. This information was circulated weekly among central and local government officers in charge of pandemic control to provide updates on the pandemic situation and to aid their decision-making on control strategies. In conclusion, for the first time in the world, we successfully demonstrated real-time estimation and prediction for the entire course of a pandemic, and which could be used routinely for planning counter-measures.